Speech to Text News and AI Updates
Follow Speech to Text developments across AI companies, labs, and open-source projects.
Latest Speech to Text news, research, benchmarks, product releases, and industry adoption updates.
Latest Articles
- NepTTS-Bench: Native-Speaker Benchmarking and NepaliMOS for Nepali Text-to-Speech — NepTTS-Bench, a Nepali text-to-speech benchmark rated by 180 native speakers, finds automated scoring favors proprietary ElevenLabs while native listeners prefer open-source Piper, because many systems carry Hindi phonology.
- Soniox STT v6 Adds Agent Text Input and Echo Filtering for Voice Agents — Soniox released STT v6, a speech-to-text API that reads what a voice agent is saying while it hears the caller, so it can catch short replies and drop the agent's own echoed voice.
- Maswao: An automatically collected, prepared, and benchmarked dataset for the South African Sign language — Researchers built Maswao, a South African Sign Language dataset assembled automatically from news broadcasts, showing a low-labour route to data for under-served sign languages, though translation scores remain low.
- GraphTFactor: A graph-based approach for proteins transcription factor prediction — GraphTFactor models protein sequences as graphs to classify transcription factors end to end, reports beating prior methods on benchmarks, and ships as open-source software with a web app.
- SecPO: Principled Adversarial Training for Prompt Injection Security — SecPO pairs a new training loss with strong adversarial training to cut prompt injection success on agents, dropping AgentDojo attack success to 8.8% versus 78.8% for Meta-SecAlign.
- Bridging the Modality Gap in Long-Form Clinical Audio: A Comparative Study of Lightweight and Heavyweight End-to-End SOAP Generation — The ASLP team's fully end-to-end system generates structured SOAP clinical notes directly from long doctor-patient audio conversations, consistently outperforming cascaded ASR+LLM baselines at both 3B and 30B parameter scales using a 1.41-million-sample training corpus.
- Building a Production Greek-English Speech Recognizer — A detailed engineering post-mortem documents building Sophea, a bilingual Greek-English ASR system: 23 training iterations, a data pipeline that recovered 88% of previously discarded Greek audio, and a 3-model ROVER ensemble that passed all 9 production quality gates with 4.26% average English WER.
- DriftSE: Speech Enhancement with Generative Drifting — DriftSE proposes a one-step generative speech enhancement framework formulated as a latent distribution equilibrium problem, using a drifting field to align noisy speech distributions to clean speech — potentially faster than multi-step diffusion approaches.
- Not All Speech Is Intent: Adaptive Self-Correcting Inference Layer for Post-ASR False Wake-Up — ASCIL is a post-ASR correction framework that cuts false wake-up errors by 54.27% relative on a session-disjoint subset, using acoustic embeddings, hesitation/silence signals, and personalized learning from past misclassifications — with under 60ms added latency.
- TokenMapper: A Step Toward Interoperable Speech Token Translation — TokenMapper enables direct token-to-token translation between structurally different speech tokenizers (single vs. multi-codebook), reducing end-to-end latency by 4.8-94.5% versus waveform bridging — accepted at AACL-IJCNLP 2026.